Why We Sometimes Reject the Evidence That We Are Loved

I sit with couples who are arguing about what happened, but beneath the facts, they are usually arguing about what the event meant.

“You knew that would hurt me.”

“I didn’t. I thought you wanted space.”

“No. You left because you didn’t care.”

The second partner tries again. “I can see why it felt like I abandoned you. I was overwhelmed, and I was afraid I would say something cruel if I stayed. I thought leaving the room would keep the argument from getting worse. I should have told you I needed twenty minutes and that I would come back.”

This is new information. It does not erase the fact that one person was left alone in pain. It does not prove that the person leaving understood every motive operating inside them. But it offers an account that was unavailable when the hurt partner first tried to explain what happened.

Sometimes the account enters the conversation and changes it. The injured partner remains hurt but becomes curious. “So you weren’t trying to punish me?”

Sometimes it never gets in.

“That is just your excuse. You left because you do not love me.”

This is one of the most painful things I watch between couples. A person is defending the belief that they were not loved, often because they are terrified of not being loved. Then evidence that they may have been loved imperfectly, misunderstood, or cared for in a way they could not recognize becomes available, and they cannot accept it.

The story was created to explain the pain. Now it is preserving the pain.

I want to name a limit at the outset, because the rest of this article depends on it. Some stories are closed because fear closed them. Others are closed because they are correct. A person whose partner has caused the same harm repeatedly is not carrying a distortion, and nothing here asks anyone to reinterpret a well-established pattern of intimidation, deception, or control as a misunderstanding. I return to this at length later. Hold it as a condition on everything in between.

The invisible sentence after the event Most relational injuries contain at least four layers:

Something happened.

We had an emotional experience of what happened.

We formed a conclusion about what it meant.

The other person had an internal experience we could not directly observe.

The first layer may be observable: You stopped responding while I was talking. The second belongs to the person who felt it: I felt dismissed and alone. The next layer is an interpretation: You stopped responding because my feelings do not matter to you. That interpretation may be accurate. It may also be partly accurate, incomplete, or wrong.

This distinction is difficult because the conclusion does not feel like a story. It feels like part of the event.

You did not answer me and you wanted me to feel unimportant become one fused fact. The observable behavior and the inferred intention become joined in the story we carry. When the other person challenges the intention, it can sound as though they are denying the behavior or invalidating the pain.

But separating those layers does not require us to question whether the pain was real. It requires us to become more precise about what we know.

I know what you did.

I know how it affected me.

I know what I concluded at the time.

I do not automatically know everything you thought, intended, understood, feared, or failed to notice.

That final category is where another person may hold information we do not have.

We do not perceive intention directly Research on couples supports this distinction. In one study, Robert Waldinger and Marc Schulz asked 156 couples to discuss an upsetting event. Participants later reviewed the interaction and reported their own intentions and emotions as well as what they believed their partners had intended.

The partners’ interpretations of intention were only weakly to moderately related to the intentions their partners reported. Relationship satisfaction explained part of the difference, and emotion helped account for the connection between satisfaction and attribution (Waldinger & Schulz, 2006).

That does not mean self-reported intention is perfect ground truth. Human beings are not fully transparent even to themselves. We can misunderstand our motives, remember ourselves favorably, or offer a cleaner explanation after the fact. But the study demonstrates something important: we do not simply detect another person’s intention. We infer it through our emotions and the condition of the relationship.

Researchers have a name for the broader version of this. Sentiment override describes the way a global feeling about a relationship can outweigh what is actually occurring in a particular moment, so that the same behavior is read as thoughtful or as hostile depending on the reservoir it lands in (Hawkins, Carrère, & Gottman, 2002).

A review of the marital attribution literature found that dissatisfied spouses are more likely than satisfied spouses to explain a partner’s behavior in ways that cast the partner negatively. These explanations may frame unwanted behavior as intentional, selfish, stable, or revealing of character. In a longitudinal study of 130 couples, Frank Fincham and Thomas Bradbury found that less benign attributions for negative partner behavior predicted lower marital satisfaction one year later, even after accounting for initial satisfaction, depression, and self-esteem (Fincham & Bradbury, 1993).

The story does not merely describe the relationship. Over time, it may help shape the relationship.

The old fear writes quickly When information is incomplete, the mind does not leave a blank space. It makes a prediction.

That word prediction is not only a metaphor. In influential models of predictive processing, the brain combines incoming information with expectations built from experience (Friston, 2010). Perception is not simply a passive recording. These models do not prove why anyone interpreted a partner negatively, but they support the broader point that expectation helps shape perception.

Someone who has repeatedly experienced abandonment may be especially quick to recognize abandonment. Someone whose needs were mocked may be alert to contempt. Someone who learned that affection can disappear without warning may watch closely for evidence that love is being withdrawn.

These sensitivities are understandable. They may have protected the person before. They may still detect dangers that other people miss. But a system trained to avoid missing a threat can also find threat in ambiguous information.

Geraldine Downey and Scott Feldman described rejection sensitivity as the tendency to anxiously expect, readily perceive, and strongly react to rejection. In their experimental work, people who expected rejection were more likely to perceive intentional rejection in ambiguous behavior. The pattern also appeared in how people interpreted insensitive behavior from romantic partners (Downey & Feldman, 1996).

Later research showed how this can become self-fulfilling. Expectations of rejection can contribute to behavior that eventually elicits more rejection from a partner. The fear influences the interpretation, the interpretation influences the response, and the response changes the relationship (Downey et al., 1998).

The loop can look like this in a couple:

I fear that I do not matter to you.

I interpret your mistake as proof that I do not matter.

I reject your explanation because it conflicts with what now feels proven.

I attack, withdraw, test you, or demand certainty you cannot provide.

You become overwhelmed or distant.

Your distance becomes more evidence that I never mattered.

This does not mean the fear created every problem. The partner may actually be inconsistent, avoidant, careless, or emotionally unavailable. What matters is whether the interpretation remains open to all the available evidence or becomes immune to anything that might revise it.

Conflict activation makes that openness harder. Acute stress narrows attention toward whatever is most immediately salient, which in an argument may be one sharp word, a facial expression, or the feared implication you do not love me — even while corrective information is being offered.

When love contradicts the self-story The paradox becomes even clearer in research on self-esteem and perceived partner regard.

Sandra Murray and colleagues found that people troubled by self-doubt underestimated how much their partners loved them. Their insecurity was not always warranted by the love and commitment their partners reported. Feeling less loved was associated with seeing the partner less positively and feeling less satisfied and optimistic about the relationship (Murray et al., 2001).

In another series of studies, people with lower self-esteem tended to resist positive feedback from romantic partners. Yet when researchers helped them reflect on the meaning and significance of a partner’s compliment, they became more able to receive it. Some effects remained two weeks later (Marigold, Holmes, & Ross, 2007).

A later randomized study extended that finding into live conflict. Denise Marigold and Joanna Anderson gave 96 dating couples a brief version of the same exercise before a conflict discussion. Without it, higher-conflict couples expected less responsiveness from their partners and subsequently reported worse outcomes. The intervention eliminated those differences (Marigold & Anderson, 2016). That does not mean expectations create every relationship outcome, or that people can think their way out of a harmful partnership. It demonstrates that what we expect from a partner can influence what becomes possible in the next interaction.

This is not merely a person refusing a compliment. It reveals how difficult it can be to absorb relational information that contradicts an established model of the self.

If I believe I am difficult to love, your love creates a conflict. I could update my view of myself and of our relationship. Or I could preserve the familiar story by deciding that you do not mean it, do not understand me, will eventually change your mind, or are trying to manipulate me.

The more painful story may still feel safer because it is known. It allows me to remain prepared. Receiving your love would require me to risk trusting information I cannot fully control.

So the same defensive system that is trying to prevent rejection may prevent reassurance from becoming emotionally real.

Why a correction may not feel corrective New information does not always replace an old impression cleanly.

In experimental research on impression revision, people have sometimes corrected what they explicitly reported while their more automatic reactions continued to reflect the original information. In one study, implicit impressions changed more when participants were able to reexamine the original behavioral information in light of what they had newly learned (Wyer, 2010).

That finding was not about couples, trauma, or attachment, so it should not be stretched into a clinical law. It does support a useful possibility: an update may require more than hearing a correction. We may need to revisit the original event and consciously reconstruct its meaning.

Brain research on impression updating should be interpreted just as carefully. A small fMRI study found that encountering behavior inconsistent with an existing impression recruited systems associated with social cognition, attention, expectancy violation, and cognitive control. Negative moral information produced especially strong updating-related responses (Mende-Siedlecki & Todorov, 2016).

There is no single “story-updating center” in the brain, and a scan cannot tell us whether one partner’s interpretation is correct. The modest lesson is that revising a person-model is active work. New information must compete with an existing prediction, and negative information about moral character may carry particular weight.

This helps explain why “I did care” may not land with the same force as “you left.” The painful event has sensory detail, emotion, bodily activation, and an immediate conclusion attached to it. The corrective information may arrive later as a sentence. It does not automatically reorganize the experience.

To dive deeper on relational concepts like these, check out: https://a.co/d/061bNz82

Updating is not erasing This is where the idea can be misused if we are not careful.

Updating the story does not mean allowing someone else to tell you what your experience was. It does not mean replacing your perception with theirs, accepting every claim of good intention, or talking yourself out of a pattern you can clearly see.

This is not an instruction for someone in a coercive or abusive relationship to reinterpret danger as love. In those relationships, explanations and claims of good intention may be part of the control itself. No one is obligated to keep reopening a conclusion that is well supported by a pattern of intimidation, deception, exploitation, or harm.

“I did not mean to hurt you” is information about stated intention. It is not an apology by itself. It does not answer whether the person was reckless, dismissive, or unwilling to consider a predictable impact. Intent and responsibility are related, but they are not identical.

A partner’s explanation becomes more credible when they can acknowledge what happened without editing it, understand the impact without arguing against it, and change the behavior that caused the injury. An explanation becomes less credible when it appears only to escape consequences, requires you to deny observable facts, or remains unchanged while the same harm repeats.

The larger pattern matters.

Imagine a partner who says, “I did not respond because I was overwhelmed, not because I wanted to punish you.” If they can recognize that disappearing for three days was harmful, develop a plan for communicating during overwhelm, and then behave differently, their explanation gains support.

If they continue disappearing, insist that your pain is the problem, and invoke overwhelm only when accountability arrives, the claimed intention tells you much less.

Updating does not require moving toward the kindest possible interpretation. It requires moving toward the interpretation that best fits all the available information.

Sometimes the update is more generous.

You were not trying to humiliate me. You became defensive and failed to notice how exposed I felt.

Sometimes it is more sobering.

I believed this was a misunderstanding. After repeated conversations and no change, I now see that my safety is not influencing your choices.

Updating is loyalty to reality, not loyalty to optimism.

There is another misuse worth naming. The easiest way to use this article is to hand it to a partner and say, “See, your story about me is wrong.” That is not updating. It is another attempt to control the meaning of someone else’s experience.

Write on Medium The harder and more honest use is reciprocal. Before asking what story your partner needs to revise, ask whether there is an impact you have refused to let revise your story about yourself. Before asking them to trust your intention, ask whether your behavior has given that intention visible form. Both people may have information to receive. I have written elsewhere about the other side of this exchange, where the person hearing feedback about their behavior cannot let it in, and I consider the two problems to be the same failure viewed from opposite chairs.

When the history is the evidence There is a version of this that I have not yet described, and it may be the most common one I see.

A person is not misreading their partner. They are reading them accurately. The partner has broken commitments, vanished during hard moments, minimized harm, or repeated the same injury after promising it would stop. This person’s vigilance is not a distortion imported from childhood. It is a summary of what has actually happened in this relationship.

For them, “consider that your story may be wrong” is poor advice, and they can usually tell. Their story is the most evidence-based thing they have. A conclusion built from a pattern is not a bias. It is a reasonable inference, and in intimate life a pattern is the strongest evidence available.

What tends to go wrong for this person is not the interpretation. It is what the accuracy of the interpretation drives them to do.

Because the fear is warranted, it becomes urgent. Urgency produces chasing: asking again, explaining more clearly, presenting more evidence, requesting a promise, pursuing a reaction that will finally prove they matter. Every attempt is understandable. Each one is an effort to make the outcome certain now rather than wait to discover it.

The difficulty is what chasing does to the evidence.

If I press you until you reassure me, I cannot interpret your reassurance. If I explain my pain until you finally respond, I do not learn whether you would have responded on your own. If I extract a promise, I have learned only that you will make a promise under pressure. Compelled care cannot be read. The behavior intended to resolve the uncertainty destroys the only information that could resolve it.

So the more useful instruction is not “trust more” or “interpret more generously.” It is closer to this: name the need clearly one time, then stop supplying the effort that has been standing in for theirs.

This is not silent withdrawal, and it is not a test administered without notice. Watching someone fail a standard they were never told about is its own form of unfairness. The need has to be stated plainly enough that they could meet it if they chose to. “When you shut down, I need you to tell me you are taking a break and when you will come back.” After that, the question is no longer how to phrase it better. The question is what they do.

Then time becomes the instrument. Not because time heals anything on its own, but because time is what separates a person’s stated intentions from their actual choices. Under pressure, people say many things. Left to choose freely, across months, they show you what they are choosing.

This posture has to be bounded, or it becomes another way of not deciding. Waiting indefinitely for someone to become different is not observation. It is hope with no expiration date, and every fresh disappointment quietly resets the clock. Decide in advance, before the next injury rather than during it, roughly what you need to see and by when. What would count as change. What would count as more of the same. What you will do when you know.

What this offers is not certainty. It offers something better than what chasing can produce: an answer you are able to trust, because you did not manufacture it.

And the update, when it comes, is often not the one people fear. It is rarely “I was wrong to be afraid.” More often it is: I was right, and I no longer have to keep proving it. I can stop working to make you into someone who chooses me, and let you show me who you are choosing to be.

The sacred space between two people I believe this is part of why relationships are so central to mental health. Other people encounter us from a position we can never occupy. They can tell us how we reach them and what becomes hard to see from inside our own defenses. We can offer them the same kind of sight.

In a relationship, each person possesses information the other cannot generate alone.

The person affected has privileged access to the impact: This is what happened inside me when you did that. This is what your behavior has come to mean because of the pattern I have lived with.

The person who acted has privileged, though imperfect, access to their internal experience: This is what I noticed. This is what I believed was happening. This is what I intended, feared, or failed to understand.

Neither person owns the whole reality.

The actor does not get to say, “Because I did not intend harm, you were not harmed.”

The receiver cannot know intention solely from impact: “Because I was harmed, you must have intended harm.”

The sacred work between them is to place both forms of information in the same room and build a fuller account than either person could build alone.

I do not use the word sacred lightly. To be told how we exist in another person’s experience is a form of relational access we cannot create for ourselves. It can be received as criticism, but it can also be received as care. The other person is showing us something about ourselves that we could not see from where we stood.

The care must remain reciprocal. The person offering feedback may also be interpreting through an old injury, a present fear, or incomplete information. Naming that possibility does not invalidate their experience. It allows the relationship to ask a more honest question: What belongs to this moment, what belongs to the history each of us brought into it, and what are we creating together now?

This kind of exchange cares for the self because neither person has to abandon their internal reality. It cares for the partner because neither person is reduced to the other’s first interpretation. And it cares for the relationship because shared reality becomes something they construct together, not something one person wins.

That account may sound like:

“You left the room without telling me whether you would return. I experienced that as abandonment, and it touched an old fear that I do not matter. I concluded that you wanted me to hurt. You are telling me that you left because you were afraid you would become cruel and believed space would protect us. I can consider that intention without pretending the way you left was okay. Next time, I need you to tell me you are taking twenty minutes and then come back when you say you will.”

Notice what this updated story does not require. It does not require the injured partner to say, “I was wrong to feel abandoned.” It does not require the other partner to confess to an intention they did not experience. It creates a shared description, a testable plan, and a path by which future behavior can build trust.

How to practice updating the story The first skill is learning to hear the invisible sentence you added after the event.

Ask yourself, What did I directly observe, and what did I conclude it meant?

“You looked at your phone while I was speaking” is observable. “You wanted me to know I was boring” is an interpretation. The interpretation may be right, but naming it as an interpretation makes it available for examination.

Then ask, What information could exist that I do not currently have? This is not an instruction to invent excuses for someone. It is an acknowledgment that another mind contains thoughts, motives, and limitations we cannot see from the outside.

The next move is to seek information rather than prosecute the conclusion:

“What was happening inside you when you picked up your phone?”

“What did you believe I needed when you left?”

“Were you aware of how that might reach me?”

“What do you understand about its impact now?”

Listen for an account, not just a denial. “I did not mean it that way” provides very little to integrate. A useful account helps reconstruct the missing internal context and still makes room for your reality.

Finally, build a both-and statement that reflects your current level of certainty:

“I know this hurt me. I am willing to consider that hurting me was not your goal.”

“I still do not understand every part of what happened, but your explanation fits some facts I had not considered.”

“I want to believe that you care. I need your behavior over time to help that become trustworthy.”

An update does not have to be complete. It can be provisional. Healthy stories are allowed to remain open while evidence accumulates.

The responsibility of the person offering the update If you want someone to revise a painful interpretation of your behavior, give them something substantial enough to work with.

Do not begin by correcting their feelings. Begin by demonstrating that you understand the event from where they stood.

“I see why my silence felt punishing.”

Then offer internal information without using it to cancel impact.

“I was flooded and shut down. I was not thinking about how the silence would affect you. That lack of awareness is part of what I need to take responsibility for.”

Finally, make the update visible in behavior.

“If I shut down again, I will tell you that I need a break, give you a time when I will return, and follow through.”

Three further obligations belong to this side of the exchange, and they are the ones most often skipped.

The first is to volunteer information before it is extracted. If your partner has to interrogate you to learn what happened inside you, the account arrives already contaminated, because neither of you can tell whether you would have offered it freely. Saying the difficult thing before being asked is worth more than saying it accurately under questioning.

The second is to expect to be tested, and to treat the testing as reasonable rather than as an insult. Someone who has been hurt repeatedly will check. Responding to the check with wounded outrage — “after everything, you still do not believe me” — converts your accountability into their offense and puts them back in the position of managing you.

The third is to stop asking when the trust will arrive. You cannot demand that another person trust your account on a timeline that suits you. If your behavior has contradicted your words for years, their existing story has considerable evidence behind it, and pressure to resolve it faster is simply more evidence that your comfort outranks their safety. Your task is not to win control of the narrative. It is to become reliable enough, for long enough, that a more accurate and safer story can eventually be built.

A relationship that can learn Sometimes our stories protect us from danger. Sometimes they protect us from disappointment. And sometimes, long after the original danger has passed, they also protect us from receiving love.

If I cannot allow any evidence that you care, you cannot succeed at caring for me in a way that becomes emotionally real. If every mistake proves rejection and every explanation is dismissed as manipulation, the relationship has no pathway back to safety. My fear of losing your love may eventually make your love impossible to experience.

This does not make me foolish, broken, or intentionally destructive. It means my protective story has become too closed to perform its original job accurately. The goal is not to become endlessly persuadable. It is to become updateable.

Healthy relationships are not relationships in which people always interpret each other correctly. That is impossible. We will misunderstand expressions, miss needs, confuse withdrawal with indifference, and mistake fear for anger. We will sometimes discover that the story we created from the outside does not match what the other person was living on the inside.

The question is whether the relationship can take in new information.

Can I tell you how your behavior reached me without claiming total knowledge of your mind?

Can you tell me what was happening inside you without using intention to erase impact?

Can we revise our understanding when the evidence warrants it?

Can we let future behavior confirm or challenge the update?

Updating the story is not surrendering reality. It is refusing to confuse our first explanation with the only possible truth.

Sometimes the update will show us that we were more loved than we knew.

Sometimes it will show us that good intentions were not enough.

Sometimes it will confirm that the pattern is unsafe and that distance is necessary.

The skill is not choosing the comforting story. The skill is allowing the story to become more accurate.

Because repair depends on more than telling the truth. It depends on whether two people can let the truth change what they thought they knew.

That willingness is one of the deepest forms of relational care I know. It says: I care about myself enough not to abandon what I experienced. I care about you enough not to confuse my interpretation with total knowledge of your mind. And I care about us enough to keep building a story that can hold more truth when more truth becomes available.

If these patterns feel familiar, R.E.A.L. Repair explores the personal beliefs, distortions, and protective responses that interfere with accountability and connection, along with practical ways to build healthier relational patterns.

References and further reading Bradbury, T. N., & Fincham, F. D. (1990). Attributions in marriage: Review and critique. Psychological Bulletin, 107(1), 3–33. Downey, G., & Feldman, S. I. (1996). Implications of rejection sensitivity for intimate relationships. Journal of Personality and Social Psychology, 70(6), 1327–1343. Downey, G., Freitas, A. L., Michaelis, B., & Khouri, H. (1998). The self-fulfilling prophecy in close relationships: Rejection sensitivity and rejection by romantic partners. Journal of Personality and Social Psychology, 75(2), 545–560. Fincham, F. D., & Bradbury, T. N. (1993). Marital satisfaction, depression, and attributions: A longitudinal analysis. Journal of Personality and Social Psychology, 64(3), 442–452. Friston, K. (2010). The free-energy principle: A unified brain theory? Nature Reviews Neuroscience, 11, 127–138. Hawkins, M. W., Carrère, S., & Gottman, J. M. (2002). Marital sentiment override: Does it influence couples’ perceptions? Journal of Marriage and Family, 64(1), 193–201. Marigold, D. C., & Anderson, J. E. (2016). Shifting expectations of partners’ responsiveness changes outcomes of conflict discussions. Personal Relationships, 23(3), 517–535. Marigold, D. C., Holmes, J. G., & Ross, M. (2007). More than words: Reframing compliments from romantic partners fosters security in low self-esteem individuals. Journal of Personality and Social Psychology, 92(2), 232–248. Mende-Siedlecki, P., & Todorov, A. (2016). Neural dissociations between meaningful and mere inconsistency in impression updating. Social Cognitive and Affective Neuroscience, 11(9), 1489–1500. Murray, S. L., Holmes, J. G., Griffin, D. W., Bellavia, G., & Rose, P. (2001). The mismeasure of love: How self-doubt contaminates relationship beliefs. Personality and Social Psychology Bulletin, 27(4), 423–436. Waldinger, R. J., & Schulz, M. S. (2006). Linking hearts and minds in couple interactions: Intentions, attributions, and overriding sentiments. Journal of Family Psychology, 20(3), 494–504. Wyer, N. A. (2010). You never get a second chance to make a first (implicit) impression: The role of elaboration in the formation and revision of implicit impressions. Social Cognition, 28(1), 1–19. The exchanges in this article are fictional composites illustrating common relational patterns. They do not describe any specific client, couple, friend, or family member.